Re: Named graphs etc

On Mar 09, 2004, at 20:20, ext Pat Hayes wrote:

>
> ... Since there are no deployed ways of distinguishing asserted from 
> non-asserted publications, it seems unlikely that the planet will now 
> decide that all the published OWL and RDF and RDFS is all just kind of 
> lying there until someone finds a way to get it asserted.
>

I expect there will be degrees of trust. Legacy RDF/OWL with no 
explicit statement
about whether it is or is not asserted will probably be trusted less 
(or not at all
for some applications) than RDF/OWL using new mechanisms to express 
assertion and
other qualifications.

As the machinery evolves, legal agreements will likely mandiate 
explicit qualification
of graphs used in interchange of knowledge falling under the terms of 
such agreements.

Over time, one could expect there will be fewer and fewer graphs with 
no qualification
whatsoever.

> RDF is a pretty weak foundation already, and we have overloaded it 
> just by putting OWL on it. It can't take any more :-)

Would constraining the interpretation of a given class of properties to
the graphs containing the statements in which they occur be overloading
RDF/OWL? Or would this rather be more of an application layer above,
but not directly impacting RDF/OWL?

> ... what is more minimal than a single XML tag?

It depends. For TriX, that is easy. And in fact, TriX *already* provides
that attribute, and has since the first published version.

But for RDF/XML, you aren't honestly proposing a change already?! ;-)

Also, a single XML attribute is easy for one serialization, but what 
about
all the other forms of expression? N3, Turtle, TriX, Squish, TriQL, 
RDFQL, RDFQ,
RDQL, etc. etc.

True, if the addition of a single syntactic flag can be proven to
provide a solid foundation for bootstrapping higher layers of trust
and authentication, then folks might be convinced to change all those
serializations, but I doubt it.

I think the vocabulary approach is both an easier sell, and also
the more "pure" RDF solution (if the semantics are doable).

>> The fact that the WG Tests are not asserted is found in the Test Cases
>> Recommendations (not that they say so explicitly).
>
> I propose that the way this will actually get done is that their 
> objective publication status is vague and underdefined, but since they 
> have no warrant, no serious agent is going to take them on trust as 
> asserted, which is all that will matter.

Right. This correlates, I think, to what I said above about degrees
of trust for non-qualified, non-authenticated graphs.

Patrick


--

Patrick Stickler
Nokia, Finland
patrick.stickler@nokia.com

Received on Wednesday, 10 March 2004 03:51:16 UTC